This paper is devoted to model-based optimization of smart well controls. Reservoir models are usually far from perfect because of the limited volume and quality of the available raw data, and the methods used to construct them, therefore model-based production optimization is extremely difficult and requires constant improvement of existing as well as the development of new approaches to its solution. The paper considers examples of some important, in our opinion, development tasks and shows possible ways of solving them, as well as a brief analysis of the results obtained with the help of approaches and methods that reflect different points of view on the uncertainty of the initial information and the accuracy of the forecast. Among the tasks considered: 1) separate and combined deployment of a smart injector and an EOR method (hot water injection); 2) use of smart wells to optimize the development of a small offshore oil field. As shown in the paper, the first task proved that quite significant synergy can arise due to the combined deployment of two IOR techniques (hot water injection and a smart injector). It also highlighted that synergy is quite insensitive to the uncertainty impact. The second task showed that the use of smart wells in combination with a proactive development strategy can significantly reduce the impact of uncertainty in the reservoir characterization on the reservoir performance. The economic efficiency of the proactive strategy in the considered example was proven to be 2-4 times higher when compared with the reactive control strategy.